Hacker News new | ask | show | jobs
by nl 5396 days ago
If Intel could move down market successfully don't you think we would have seen a reasonable IA facsimile of an iPhone or an iPad after all these years?

No, chip design doesn't work that quickly. Intel has accelerated their development program for Atom[1], but at the moment Intel doesn't have a processor that competes at the power usage of current smart phones. Atom is supposed to get there in 2012 - they are usually pretty reliable (see Intel's Tick/Tock strategy[2]), but we'll see.

They are planning to do 22nm Atoms in 2013, and 14nm in 2014. TSMC is supposed to start shipping ARM on 28nm in Q4 2011[3], but historically they are usually late. TSMC is allegedly planning a 20nm fab, but it is unclear how far off that is.

Nokia (remember them?) was signed up to do smartphones on Atom in 2012, but then they switched to WinMo (which is ARM only).

I personally don't think they'll do well in smartphone, but tablets is a whole other market.

ARM,as a mobile optimized architecture, will never compete on performance grounds with IA

That's not what nVidia think[4]. There are 32/64bit issues with ARM, but beyond that there is little in the architecture itself that means it can't compete with IA.

[1] http://www.eetimes.com/electronics-news/4216089/Intel-rewrit...

[2] http://en.wikipedia.org/wiki/Intel_Tick-Tock

[3] http://www.xbitlabs.com/news/other/display/20110201212245_TS...

[4] http://www.engadget.com/2011/01/05/nvidia-announces-project-...

1 comments

The problem for Intel is not so much that they can't make a competitive chip, but that they can't make one at margins that will sustain their fabs.
I think it's premature to say that.

Intel has the best fabs in the world, and they can make chips as cheaply as anyone.

It is possible that at some point in the future Intel's margins won't sustain the continual investment in new fabs. That's a fair way off at the moment though.